Satellite-based Phenology Analysis in Evaluating the Response of Puerto Rico and the US Virgin Island’s Tropical Forests to the 2017 Hurricanes

Abstract:

The functionality of tropical forest ecosystems and their productivity is highly related to the timing of phenological events. Understanding forest responses to major climate events can aid in identifying the role diversity plays in forest resilience. This research will utilize Landsat satellite data and ground-based Forest Inventory and Analysis (FIA) data to investigate Puerto Rico and the US Virgin Island’s (PRVI) tropical forests after 2017s two major hurricanes and how forest structure and composition affected resilience. Harmonizing these two datasets also allows us to validate the remote sensing methodology and investigate whether this is an accurate approach for estimating forest health in areas lacking in-situ data. We will perform extensive cloud masking processes on the satellite imagery to produce repaired, near cloud-free imagery, which will then be used to produce annual phenology curves. FIA data will be used to estimate tree mortality, tree damage, change in aboveground live biomass (AGLB), and species diversity. Regression and spatial autoregressive models will be conducted to explore the relationship between the FIA data and phenology anomalies and to analyze and compare landscape trends. We expect to see significant changes in phenology trends and species diversity from before and after the hurricanes. We also anticipate a decrease in overall forest health and AGLB, and higher resilience in areas with higher species diversity. Results and findings will help construct a base understanding of PRVI’s tropical forests dynamic relative to climate change and give a clearer indication of forest vulnerability. Furthermore, this research will demonstrate approaches and techniques that can be further applied to larger, global sustainability goals to reduce biodiversity loss and sustain living systems in times of climate variability and change.